Optimizing cluster sizing using a real world benchmark

[rodrigo-lourenco-lopes]

This PR adds a new chaos day esperiment: "Optimizing cluster sizing using a real world benchmark".

Additionally, it fixes some links and typos in the readme.md.

[rodrigo-lourenco-lopes]

Hey @npepinpe this is still a draft and I have still to update the benchmark values to reflect the changes that were made in the gateway by adding more memory, among some other improvements to the text, but I just wanted to ask before some things that we should include or not.

• Should we add as a background and motivation the Saas advanced Offering topic?
• Should we add the cost reduction as a motivation, or is that something that might not be ideal for external reading? And if the answer yes should we also display some metrics on cost reduction?
• Is it too much to include all the resources of out clusters as we currently have laid out in the tables?

[npepinpe]

Should we add as a background and motivation the Saas advanced Offering topic?

I don't think this is relevant for most people reading this.

Should we add the cost reduction as a motivation, or is that something that might not be ideal for external reading? And if the answer yes should we also display some metrics on cost reduction?

Same, I think we don't need to mention that.

[npepinpe]

🚀 Generally looks good

I would remove the link to accounts though, this is private so not relevant for most people.

[ChrisKujawa]

TBH I personally would separate this from this blog post. As the idea for this and work came from the Camunda Exporter Project and was then reused for this project as well.

What you could do here is to say we used a different way of benchmark with more realistic workload. You will not go into more details as this deserves an own blog post, which is upcoming.

[rodrigo-lourenco-lopes]

Ups, I did not know this was an upcoming post, should we cut this completely, or still use as motivation and link it to the upcoming post?

[rodrigo-lourenco-lopes]

Now with the new values for the gateway with more memory, I changed the original motivation of 40 tasks/s to 100 to be easier to explain in this post. Instead of trying to explain all the iterations of benchmarking we went through.
@npepinpe @Zelldon does this make sense to you?